Research on TCN Model Based on Improved Random Feature Selection in Email Filtering
ZHANG Wei
DENG Shijie
YU Guibo
Abstract：In recent years,instant messaging software has developed rapidly and become the main information tool in people's daily social life.However,in the work environment,email is still considered a more formal form of communication and is mainly used to send formal reports and notifications.However,spam emails containing advertising,violence,pornography and even Trojan viruses take up our mailbox resources and affect its reliability,causing inconvenience to our work.Therefore,this paper investigates a temporal convolutional neural network(TCN)model based on improved random forest feature selection to accurately identify the content and information of spam emails.Based on the TCN classification model,the model incorporates the Random Forest(RF)with the sparrow optimisation algorithm(SSA)for dimensionality reduction of the original features,which is used to remove poorly correlated and unimportant features and retain effective features with a certain contribution rate to obtain the optimal subset of fea-tures.In order to verify the reliability of the model,comparative experiments are carried out on the Spambase dataset at UCI to anal-yse and evaluate the model based on five indices,which are F1 score,precision,recall,accuracy and number of training parame-ters.The experiments show that the model can effectively identify spam emails,and the optimal feature subset improves the recogni-tion accuracy while reducing the complexity of the model,and the evaluation metrics are all better than the experimental comparison model.
Keywords：spam emailsSSARFfeature selectiondimensionality reductionTCN
Publication Date：2025-12-20
Online Publishing Date：2026-03-23(First online date of this platform, not the publication date of the document)
Pages：7( 102-107,159 )